W3C home > Mailing lists > Public > xml-editor@w3.org > April to June 2000

Re: Official ISO 639 changes

From: John Cowan <cowan@locke.ccil.org>
Date: Tue, 25 Apr 100 18:47:34 -0400 (EDT)
Message-Id: <200004252247.SAA13008@locke.ccil.org>
To: elharo@metalab.unc.edu (Elliotte Rusty Harold)
Cc: unicode@unicode.org, xml-dev@xml.org, xml-editor@w3.org
Elliotte Rusty Harold scripsit:

> Whether it's important or not, all XML 1.0 conforming parsers today 
> do check xml:lang values, at least to the extent of making sure 
> they're two letters and not three. Failing to do so is a 
> well-formedness error because of Production 35.

Oddly enough, this is not true.  Production 35 is not reachable from any
other production, and there is no "must" language in clause 2.12
(except "must be declared in valid documents", which is not relevant).
A document meets the technical definition of well-formedness in 
clause 2.1 even if its xml:lang attribute values are crud.

One view is that conforming to production 35 (or, equivalently, to the
rules in RFC 1766) should be made a validity (not well-formedness)
constraint.  The other view is that productions 35-38 never were
normative and should be disposed of in favor of motherhood language,
something like "xml:lang attribute values should conform to RFC 1766
or its successor."

No decision has been reached.

John Cowan                                   cowan@ccil.org
       I am a member of a civilization. --David Brin
Received on Tuesday, 25 April 2000 18:34:59 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:37:39 UTC